Clustering and Distance Metrics in Biology

The video tutorial explains hierarchical clustering, a method used to order data based on similarity, often visualized with heat maps. It covers the process of clustering genes, using examples to illustrate the steps involved. The tutorial also discusses different distance metrics like Euclidean and Manhattan distances, and methods for comparing clusters, such as centroid, single linkage, and complete linkage. The video concludes with a summary of the key concepts and encourages viewers to explore further.

What is hierarchical clustering often associated with?

• Bar charts

• Heat maps

• Line graphs

• Pie charts

In a heat map, what do the columns typically represent?

• Different colors

• Different expressions

• Different samples

• Different genes

What color typically signifies high expression of a gene in a heat map?

• Green

• Blue

• Red

• Purple

What is the first step in clustering genes?

• Identify the most similar gene to a given gene

• Merge all genes into one cluster

• Determine the color of each gene

• Calculate the average expression

Which genes are merged into the first cluster in the example?

• Genes 2 and 4

• Genes 1 and 2

• Genes 1 and 3

• Genes 3 and 4

What does a dendrogram indicate in hierarchical clustering?

• The number of genes

• The expression level of samples

• The similarity and order of clusters

• The color of each gene

What is the Euclidean distance used for in clustering?

• To determine the color of genes

• To calculate the average expression

• To measure similarity between genes

• To reorder the columns

How is the Manhattan distance different from the Euclidean distance?

• It uses absolute values of differences

• It averages the differences

• It multiplies the differences

• It uses the square root of differences

What is one method to compare clusters?

• Using the product of each cluster

• Using the median of each cluster

• Using the average of each cluster

• Using the sum of each cluster

What is the default setting for the hclust function in R?

• Complete linkage

• Single linkage

• Average linkage

• Centroid linkage